Crate facet_deserialize

Expand description

Logo by Misiasart

Thanks to all individual and corporate sponsors, without whom this work could not exist:

An event-based approach to facet-deserialize

§License

Licensed under either of:

Apache License, Version 2.0 (LICENSE-APACHE or http://www.apache.org/licenses/LICENSE-2.0)
MIT license (LICENSE-MIT or http://opensource.org/licenses/MIT)

at your option.

Structs§

DeserError: A JSON parse error, with context. Never would’ve guessed huh.
DeserErrorMessage: A wrapper type for displaying deser error messages
NextData: Carries the current parsing state and the in-progress value during deserialization. This bundles the mutable context that must be threaded through parsing steps.
Span: A span in the input, with a start position and length
Spanned: A value of type T annotated with its Span
Subspan: A Subspan variant of a Span
Substack: Container for subspans based on span type

Enums§

Cooked: A Cooked variant of a Span (byte indexed)
DeserErrorKind: An error kind for JSON parsing.
Expectation: Expected next input token or structure during deserialization.
Instruction: Instructions guiding the parsing flow, indicating the next expected action or token.
Outcome: Outcome of parsing the next input element.
PopReason: Reasons for popping a state from the stack, indicating why a scope is ended.
Raw: A Raw variant of a Span (format-specific index)
Scalar: A scalar value used during deserialization. u64 and i64 are separated because i64 doesn’t fit in u64, but having u64 is a fast path for 64-bit architectures — no need to go through u128 / i128 for everything
SubspanMeta: Metadata about a subspan, providing context for how the subspan relates to the parent span or other subspans.
ValueReason: Reasons for expecting a value, reflecting the current parse context.

Traits§

Format: Trait defining a deserialization format. Provides the next parsing step based on current state and expected input.
InputDebug: Trait for handling input data in error reporting and debugging. Provides methods to convert input slices to human-readable strings and to create byte arrays for error storage.
Spannable: Trait for types that can be annotated with a Span.
SubstackBehavior: This trait allows the compiler to optimize away Substack-related code for formats with span types that don’t use subspans, making it zero-cost.
ToCooked: Trait handling conversion regardless of Format::SpanType to Span<Cooked>

Functions§

deserialize: Deserialize a value of type T from raw input bytes using format F.
deserialize_wip: Deserializes a working-in-progress value into a fully materialized heap value. This function drives the parsing loop until the entire input is consumed and the value is complete.

Type Aliases§

NextResult: The result of advancing the parser: updated state and parse outcome or error.
Pos: Position in the input (byte index)